Re: [PATCH 1/1] usb: gadget: f_uac2: fix error handling at afunc_bind

[Peter Chen <hzpeterchen@gmail.com>]

On Thu, Nov 03, 2016 at 10:50:23AM +0200, Felipe Balbi wrote:
> 
> Hi,
> 
> Peter Chen <peter.chen@nxp.com> writes:
> > The current error handling flow uses incorrect goto label, fix it
> >
> > Cc: <stable@vger.kernel.org>
> > Fixes: d12a8727171c ("usb: gadget: function: Remove
> > 	redundant usb_free_all_descriptors")
> > Signed-off-by: Peter Chen <peter.chen@nxp.com>
> 
> seems like there's more than one fix here.
> 
> > ---
> >  drivers/usb/gadget/function/f_uac2.c | 20 +++++++++++---------
> >  1 file changed, 11 insertions(+), 9 deletions(-)
> >
> > diff --git a/drivers/usb/gadget/function/f_uac2.c b/drivers/usb/gadget/function/f_uac2.c
> > index cd214ec8..3f4e478 100644
> > --- a/drivers/usb/gadget/function/f_uac2.c
> > +++ b/drivers/usb/gadget/function/f_uac2.c
> > @@ -1067,13 +1067,13 @@ afunc_bind(struct usb_configuration *cfg, struct usb_function *fn)
> >  	agdev->out_ep = usb_ep_autoconfig(gadget, &fs_epout_desc);
> >  	if (!agdev->out_ep) {
> >  		dev_err(dev, "%s:%d Error!\n", __func__, __LINE__);
> > -		goto err;
> > +		return ret;
> >  	}
> >  
> >  	agdev->in_ep = usb_ep_autoconfig(gadget, &fs_epin_desc);
> >  	if (!agdev->in_ep) {
> >  		dev_err(dev, "%s:%d Error!\n", __func__, __LINE__);
> > -		goto err;
> > +		return ret;
> >  	}
> >  
> >  	uac2->p_prm.uac2 = uac2;
> > @@ -1091,13 +1091,14 @@ afunc_bind(struct usb_configuration *cfg, struct usb_function *fn)
> >  	ret = usb_assign_descriptors(fn, fs_audio_desc, hs_audio_desc, NULL,
> >  				     NULL);
> >  	if (ret)
> > -		goto err;
> > +		return ret;
> 
> this is one fix
> 
> >  
> >  	prm = &agdev->uac2.c_prm;
> >  	prm->max_psize = hs_epout_desc.wMaxPacketSize;
> >  	prm->rbuf = kzalloc(prm->max_psize * USB_XFERS, GFP_KERNEL);
> >  	if (!prm->rbuf) {
> >  		prm->max_psize = 0;
> > +		ret = -ENOMEM;
> 
> but initializing ret to -ENOMEM is a separate fix altogether. 
> 

Ok, I will have a bug-fix just fixing incorrect label, and have 
another patch to change return value for -next.

-- 

Best Regards,
Peter Chen